#lang scribble/doc
@(require 
  scribble/manual
  (for-label racket
             picturing-programs/main
             ;picturing-programs/io-stuff
	     ;picturing-programs/tiles
             ;picturing-programs/dummy
	     ; picturing-programs/map-image
             2htdp/image
             teachpack/2htdp/universe
             (only-in lang/htdp-beginner check-expect)
             ))

@; teachpack["picturing-programs"]{Picturing Programs}
@title{@italic{Picturing Programs} Teachpack}
@author{Stephen Bloch}

@defmodule[picturing-programs]

@section{About This Teachpack}

@;Testing, testing: @racket[(list 'testing 1 2 3)].
@;
@;This is a reference to the @racket[list] function (which is a nice link).
@;Now a reference to @racket[triangle] (good link),
@;and @racket[big-bang] (good link),
@;and @racket[show-it] (good link),
@;and @racket[crop-top] (underlined in red, not a link), 
@;and @racket[map-image] (underlined in red, not a link),
@;and @racket[dummyvar] (how does this look?),
@;and @racket[with-input-from-url] (underlined in red, not a link), 
@;which are defined in several different places.

Provides a variety of functions for combining and manipulating images
and running interactive animations.
It's intended to be used with the textbook
@hyperlink["http://www.picturingprograms.com" "Picturing Programs"].

@section{Installation}
This package should be bundled with DrRacket version 5.1 and later, so there should be
no installation procedure.

@section{Functions from @racketmodname[2htdp/image] and @racketmodname[2htdp/universe]}

This package includes all of 
@racketmodlink[2htdp/image]{the image teachpack} and
and
@racketmodlink[2htdp/universe]{the universe teachpack},
so if you're using this teachpack, @italic{don't} also load either of those.
See the above links for how to use those teachpacks.

It also supersedes the older @racket[tiles] and @racket[sb-world] teachpacks,
so if you have those, don't load them either; use this instead.

This package also provides the following additional functions:

@; @include-section{image.rkt}

@section{Animation support}
Since the 
@hyperlink["http://www.picturingprograms.com" "Picturing Programs"]
textbook introduces animations with image models before other model
types, we provide a draw handler for the simple case in which the
model is exactly what should be displayed in the animation window:

@defproc[(show-it [img image?]) 
         image?]{Returns the given image unaltered.  Useful as a draw handler for animations whose model is an image.}

@section{New image functions}
@; @defmodule*/no-declare[(picturing-programs/tiles)]
@declare-exporting[picturing-programs/private/tiles picturing-programs]

@defproc[(rotate-cw [img image?])
         image?]{Rotates an image 90 degrees clockwise.}

@defproc[(rotate-ccw [img image?])
         image?]{Rotates an image 90 degrees counterclockwise.}

@defproc[(rotate-180 [img image?])
         image?]{Rotates an image 180 degrees around its center.}

@defproc[(crop-top [img image?] [pixels natural-number/c])
         image?]{Chops off the specified number of pixels from the top of the image.}

@defproc[(crop-bottom [img image?] [pixels natural-number/c])
         image?]{Chops off the specified number of pixels from the bottom of the image.}

@defproc[(crop-left [img image?] [pixels natural-number/c]) 
         image?]{Chops off the specified number of pixels from the left side of the image.}

@defproc[(crop-right [img image?] [pixels natural-number/c]) 
         image?]{Chops off the specified number of pixels from the right side of the image.}

@defproc[(flip-main [img image?])
 	 image?]{Reflects an image across the line x=y, moving the pixel
at coordinates (x,y) to (y,x).  The top-right corner becomes the
bottom-left corner, and vice versa.  Width and height are swapped.}

@defproc[(flip-other [img image?])
	 image?]{Reflects an image by moving the pixel at coordinates
(x,y) to (h-y, w-x).  The top-left corner becomes the bottom-right
corner, and vice versa.  Width and height are swapped.}

@defproc[(reflect-vert [img image?]) 
         image?]{The same as @racket[flip-vertical]; retained for compatibility.}

@defproc[(reflect-horiz [img image?]) 
         image?]{The same as @racket[flip-horizontal]; retained for compatibility.} 

@defproc[(reflect-main-diag [img image?])
	 image?]{The same as @racket[flip-main]; retained for
compatibility.}

@defproc[(reflect-other-diag [img image?])
	 image?]{The same as @racket[flip-other]; retained for
compatibility.}

@section{Variables}

@; @defmodule*/no-declare[(picturing-programs/book-pictures)]
@declare-exporting[picturing-programs/private/book-pictures picturing-programs]

This teachpack also defines variable names for some of the pictures used in the textbook.

@defthing[pic:bloch image?]{A picture of the author, c. 2005.}
@defthing[pic:hieroglyphics image?]{A picture of a stone tablet with
hieroglyphics on it.}
@defthing[pic:hacker image?]{A picture of a student sitting at a
computer.}
@defthing[pic:book image?]{A picture of a book with a question mark.}
@defthing[pic:stick-figure image?]{A picture of a stick figure, built
from geometric primitives.}
@defthing[pic:scheme-logo image?]{A picture of a DrScheme/DrRacket
logo.}
@defthing[pic:calendar image?]{A picture of an appointment calendar.}

Note that these seven variable names happen to start with "pic:", to
distinguish them from anything you might define that happens to be named
"calendar" or "book", but you can name a variable anything you want; in
particular, there's no requirement that your names start with "pic:".

@section{Pixel functions}

@; @defmodule*/no-declare[(picturing-programs/map-image)]
@declare-exporting[picturing-programs/private/map-image picturing-programs]


The above functions allow you to operate on a picture as a whole, but sometimes
you want to manipulate a picture pixel-by-pixel.

@subsection{Colors and pixels}

Each pixel of a bitmap image has a @racket[color], a built-in structure with
four components --  red, green, blue, and alpha -- each represented by an
integer from 0 to 255.  Larger alpha values are "more opaque": an image with
alpha=255 is completely opaque, and one with alpha=0 is completely
transparent.

Even if you're not trying to get transparency effects, alpha is also used
for dithering to smooth out jagged edges.  In
@racket[(circle 50 "solid" "red")], the pixels inside the circle are pure
red, with alpha=255; the pixels outside the circle are transparent (alpha=0);
and the pixels on the boundary are red with various alpha values (for example,
if one quarter of a pixel's area is inside
the mathematical boundary of the circle, that pixel's alpha value will be
63).

@defproc[(name->color [name (or/c string? symbol?)])
         (or/c color? false/c)]{
Given a color name like "red", 'turquoise, "forest green", @italic{etc.}, returns the corresponding
color struct, showing the red, green, blue, and alpha components.  If the name isn't
recognized, returns @racket[false].}

@defproc[(colorize [thing (or/c color? string? symbol? false/c)])
	(or/c color? false/c)]{
Similar to @racket[name->color], but accepts colors and @racket[false] as
well: colors produce themselves, while @racket[false] produces a transparent
color.}

@defproc[(color=? [c1 (or/c color? string? symbol? false/c)]
		  [c2 (or/c color? string? symbol? false/c)])
	 boolean?]{
Compares two colors for equality.  As with @racket[colorize], treats
@racket[false] as a transparent color (i.e. with an alpha-component of 0).
All colors with alpha=0 are considered equal to one another, even if they have
different red, green, or blue components.}

@defproc[(get-pixel-color [x natural-number/c] [y natural-number/c] [pic image?])
         color?]{

Gets the color of a specified pixel in the given image.  If x and/or y are outside
the bounds of the image, returns a transparent color.}

@subsection{Specifying the color of each pixel of an image}

@defproc[(build-image [width natural-number/c]
                      [height natural-number/c]
                      [f (-> natural-number/c natural-number/c color?)])
         image?]{

Builds an image of the specified size and shape by calling the specified
function on the coordinates of each pixel.  For example,
@codeblock|{
; fuzz : image -> image
(define (fuzz pic)
  (local [; near-pixel : num(x) num(y) -> color
          (define (near-pixel x y)
            (get-pixel-color (+ x -3 (random 7))
                             (+ y -3 (random 7))
                             pic))]
     (build-image (image-width pic)
                  (image-height pic)
                  near-pixel)))
}|
produces a fuzzy version of the given picture by replacing each pixel with a
randomly chosen pixel near it.}

@defproc[(build-image/extra
         [width natural-number/c]
         [height natural-number/c]
         [f (-> natural-number/c natural-number/c any/c color?)] [extra any/c]) image?]{
Passes the @racket[extra] argument in as a third argument in each call
to @racket[f].  This allows students who haven't learned closures yet to do pixel-by-pixel image
manipulations inside a function depending on a parameter of that function.

For example, the above @racket[fuzz] example could also be written as
@codeblock|{
; near-pixel : number(x) number(y) image -> color
(define (near-pixel x y pic)
  (get-pixel-color (+ x -3 (random 7))
                   (+ y -3 (random 7))
                   pic))
; fuzz : image -> image
(define (fuzz pic)
  (build-image/extra (image-width pic)
                     (image-height pic)
                     near-pixel
                     pic))
}|
}

@defproc[(build4-image [width natural-number/c] [height natural-number/c]
		       [red-function (-> natural-number/c natural-number/c natural-number/c)]
		       [green-function (-> natural-number/c natural-number/c natural-number/c)]
		       [blue-function (-> natural-number/c natural-number/c natural-number/c)]
		       [alpha-function (-> natural-number/c natural-number/c
natural-number/c)])
		image?]{
A version of @racket[build-image] for students who don't know about structs yet.
Each of the four functions takes in the x and y coordinates of a pixel, and
should return an integer from 0 through 255 to determine that color component.}

@defproc[(build3-image [width natural-number/c] [height natural-number/c]
                       [red-function (-> natural-number/c natural-number/c natural-number/c)]
                       [green-function (-> natural-number/c natural-number/c natural-number/c)]
                       [blue-function (-> natural-number/c natural-number/c natural-number/c)])
         image?]{
Just like @racket[build4-image], but without specifying the alpha component
(which defaults to 255, fully opaque).}

@defproc*[([(map-image [f (-> color? color?)] [img image?]) image?]
           [(map-image [f (-> natural-number/c natural-number/c color?  color?)] [img image?]) image?])]{
Applies the given function to each pixel in a given image, producing a
new image the same size and shape.  The color of each pixel in the
result is the result of calling f on the corresponding
pixel in the input.  If f accepts 3 parameters, it will be given the x
and y coordinates and the color of the old pixel; if it accepts 1, it
will be given only the color of the old pixel.

An example with a 1-parameter function:
@codeblock|{
; lose-red : color -> color
(define (lose-red old-color)
  (make-color 0 (color-green old-color) (color-blue old-color)))

(map-image lose-red my-picture)}|
produces a copy of @racket[my-picture] with all the red leached out,
leaving only the blue and green components.

Since @racket[make-color] defaults alpha to 255,
this definition of @racket[lose-red] discards any alpha information (including edge-dithering)
that was in the original image.  To preserve this information, one could write
@racketblock[
(define (lose-red-but-not-alpha old-color)
  (make-color 0 (color-green old-color) (color-blue old-color) (color-alpha
old-color)))]

An example with a 3-parameter (location-sensitive) function:
@codeblock|{
; apply-gradient : num(x) num(y) color -> color
(define (apply-gradient x y old-color)
  (make-color (min (* 3 x) 255)
              (color-green old-color)
              (color-blue old-color)))

(map-image apply-gradient my-picture)}|
produces a picture the size of @racket[my-picture]'s bounding rectangle,
replacing the red component with a smooth color gradient increasing
from left to right, but with the green and blue components unchanged.}

@defproc*[([(map-image/extra [f (-> color? any/c color?)] [img image?] [extra any/c]) image?]
	  [(map-image/extra [f (-> natural-number/c natural-number/c color? any/c color?)] [img image?] [extra any/c]) image?])]{
Passes the @racket[extra] argument in as an additional argument in each call
to @racket[f].  This allows students who haven't learned closures yet to do pixel-by-pixel image
manipulations inside a function depending on a parameter of that function.

For example,
@codeblock|{
; clip-color : color number -> color
(check-expect (clip-color (make-color 30 60 90) 100)
	      (make-color 30 60 90))
(check-expect (clip-color (make-color 30 60 90) 50)
	      (make-color 30 50 50))
(define (clip-color c limit)
    (make-color (min limit (color-red c))
		(min limit (color-green c))
		(min limit (color-blue c))))

; clip-picture-colors : number(limit) image -> image
(define (clip-picture-colors limit pic)
    (map-image/extra clip-color pic limit))
}|

This @racket[clip-picture-colors] function clips each of the
color components at most to the specified limit.

Another example, using x and y coordinates as well:
@codeblock|{
; new-pixel : number(x) number(y) color height -> color
(check-expect (new-pixel 36 100 (make-color 30 60 90) 100)
              (make-color 30 60 255))
(check-expect (new-pixel 58 40 (make-color 30 60 90) 100)
              (make-color 30 60 102))
(define (new-pixel x y c h)
  (make-color (color-red c)
              (color-green c)
              (real->int (* 255 (/ y h)))))

; apply-blue-gradient : image -> image
(define (apply-blue-gradient pic)
  (map-image/extra new-pixel pic (image-height pic)))
}|
This @racket[apply-blue-gradient] function changes the blue component of an image to increase gradually
from the top to the bottom of the image, (almost) reaching 255 at the bottom of the image.
}

@defproc[(map4-image 
[red-func (-> natural-number/c natural-number/c natural-number/c natural-number/c natural-number/c natural-number/c natural-number/c)]
[green-func (-> natural-number/c natural-number/c natural-number/c natural-number/c natural-number/c natural-number/c natural-number/c)]
[blue-func (-> natural-number/c natural-number/c natural-number/c natural-number/c natural-number/c natural-number/c natural-number/c)]
[alpha-func (-> natural-number/c natural-number/c natural-number/c natural-number/c natural-number/c natural-number/c natural-number/c)]
[img image?])
image?]{

A version of map-image for students who don't know about structs yet.  Each of the
four given functions is assumed to have the contract
@racketblock[num(x) num(y) num(r) num(g) num(b) num(alpha) -> num]
For each pixel in the original picture, applies the four
functions to the x coordinate, y coordinate, red, green, blue, and alpha
components of the pixel.
The results of the four functions are used as the red, green, blue, and alpha
components in the corresponding pixel of the resulting picture.

For example,
@codeblock|{
; each function : num(x) num(y) num(r) num(g) num(b) num(a) -> num
(define (zero x y r g b a) 0)
(define (same-g x y r g b a) g)
(define (same-b x y r g b a) b)
(define (same-alpha x y r g b a) a)
(map4-image zero same-g same-b same-alpha my-picture)}|
produces a copy of @racket[my-picture] with all the red leached out,
leaving only the blue, green, and alpha components.

@codeblock|{
; each function : num(x) num(y) num(r) num(g) num(b) num(a) -> num
(define (3x x y r g b a) (min (* 3 x) 255))
(define (3y x y r g b a) (min (* 3 y) 255))
(define (return-255 x y r g b a) 255)
(map4-image 3x zero 3y return-255 my-picture)}|
produces an opaque picture the size of @racket[my-picture]'s bounding rectangle,
with a smooth color gradient with red increasing from left to
right and blue increasing from top to bottom.
}

@defproc[(map3-image
[red-func (-> natural-number/c natural-number/c natural-number/c natural-number/c natural-number/c natural-number/c)]
[green-func (-> natural-number/c natural-number/c natural-number/c natural-number/c natural-number/c natural-number/c)]
[blue-func (-> natural-number/c natural-number/c natural-number/c natural-number/c natural-number/c natural-number/c)]
[img image?])
image?]{
Like @racket[map4-image], but not specifying the alpha component.  Note that
the red, green, and blue functions also @italic{don't take in} alpha values.
Each of the three given functions is assumed to have the contract
@racketblock[num(x) num(y) num(r) num(g) num(b) -> num]
For each pixel in the original picture, applies the three functions to the x
coordinate, y coordinate, red, green, and blue components of the pixel.
The results are used as a the red, green, and blue components in the
corresponding pixel of the resulting picture.

The alpha component in the resulting picture is copied from the source
picture.  For example,
@codeblock|{
; each function : num(x) num(y) num(r) num(g) num(b) -> num
(define (zero x y r g b) 0)
(define (same-g x y r g b) g)
(define (same-b x y r g b) b)
(map3-image zero same-g same-b my-picture)}|
produces a copy of @racket[my-picture] with all the red leached out; parts of
the picture that were transparent are still transparent, and parts that were
dithered are still dithered.
@codeblock|{
; each function : num(x) num(y) num(r) num(g) num(b) num(a) -> num
(define (3x x y r g b a) (min (* 3 x) 255))
(define (3y x y r g b a) (min (* 3 y) 255))
(map3-image zero 3x 3y my-picture)}|
produces a @racket[my-picture]-shaped "window" on a color-gradient.
}

@defproc*[([(fold-image [f (-> color? any/c any/c)] [init any/c] [img image?]) any/c]
           [(fold-image [f (-> natural-number/c natural-number/c color? any/c any/c)] [init any/c] [img image?]) any/c])]{
Summarizes information from all the pixels of an image.
The result is computed by applying f successively to each pixel, starting with @racket[init].
If @racket[f] accepts four parameters, it is called with the coordinates and color of each
pixel as well as the previously-accumulated result; if it accepts two parameters, it is
given just the color of each pixel and the previously-accumulated result.
You may not assume anything about the order in which the pixels are visited, only
that each pixel will be visited exactly once.

An example with a 2-parameter function:
@codeblock|{
; another-white : color number -> number
(define (another-white c old-total)
   (+ old (if (color=? c "white") 1 0)))

; count-white-pixels : image -> number
(define (count-white-pixels pic)
   (fold-image another-white 0 pic))}|

Note that the accumulator isn't restricted to be a number: it could be a structure or a list,
enabling you to compute the average color, or a histogram of colors, etc.
}

@defproc*[([(fold-image/extra [f (-> color? any/c any/c any/c)] [init any/c] [img image?] [extra any/c]) any/c]
	  [(fold-image/extra [f (-> natural-number/c natural-number/c color? any/c any/c any/c)] [init any/c] [img image?] [extra any/c]) any/c])]{
Like @racket[fold-image], but passes the @racket[extra] argument in as an additional argument in each call
to @racket[f].  This allows students who haven't learned closures yet to call @racket[fold-image] on an
operation that depends on a parameter to a containing function.

For example,
@codeblock|{
; another-of-color : color number color -> number
(define (another-of-color c old color-to-count)
   (+ old (if (color=? c color-to-count) 1 0)))

; count-pixels-of-color : image color -> number
(define (count-pixels-of-color pic color-to-count)
   (fold-image/extra count-pixels-of-color 0 pic))
}|
}

@defproc[(real->int [num real?])
         integer?]{
Not specific to colors, but useful if you're building colors by arithmetic.
For example,
@codeblock|{
; bad-gradient : num(x) num(y) -> color
(define (bad-gradient x y)
  (make-color (* 2.5 x) (* 1.6 y) 0))
(build-image 50 30 bad-gradient)

; good-gradient : num(x) num(y) -> color
(define (good-gradient x y)
  (make-color (real->int (* 2.5 x)) (real->int (* 1.6 y)) 0))
(build-image 50 30 good-gradient)
}|
The version using @racket[bad-gradient] crashes because color components must be exact integers.
The version using @racket[good-gradient] works.}


@section{Input and Output}

@; @defmodule*/no-declare[(picturing-programs/io-stuff)]
@declare-exporting[picturing-programs/private/io-stuff picturing-programs]


This teachpack also provides several functions to help in testing
I/O functions (in Advanced Student language; ignore this section if
you're in a Beginner or Intermediate language):

@defproc[(with-input-from-string [input string?]
                                 [thunk (-> any/c)])
         any/c]{

Calls @tt{thunk}, which presumably uses @racket[read],
in such a way that @racket[read] reads from @tt{input} rather than from
the keyboard.}

@defproc[(with-output-to-string [thunk (-> any/c)])
         string?]{

Calls @tt{thunk}, which presumably uses @racket[display], @racket[print],
@racket[write], and/or @racket[printf], in such a way that its output is
accumlated into a string, which is then returned.}

@defproc[(with-input-from-file [filename string?]
           [thunk (-> any/c)]) any/c]{
Calls @tt{thunk}, which presumably uses @racket[read],
in such a way that @racket[read] reads from the specified file
rather than from the keyboard.}

@defproc[(with-output-to-file (filename string?) (thunk (-> any/c))) any/c]{
Calls @tt{thunk}, which presumably uses @racket[display], @racket[print],
@racket[write], and/or @racket[printf], in such a way that its output is
redirected into the specified file.}

@defproc[(with-input-from-url (url string?) (thunk (-> any/c))) any/c]{
Calls @tt{thunk}, which presumably uses @racket[read],
in such a way that @racket[read] reads from the HTML source of the
Web page at the specified URL rather than from the keyboard.}

@defproc[(with-io-strings (input string?) (thunk (-> any/c))) string?]{
Combines @racket[with-input-from-string] and @racket[with-output-to-string]:
calls @tt{thunk} with its input coming from @tt{input} and accumulates
its output into a string, which is returned.  Especially useful for testing:
@codeblock|{
; ask : string -> prints output, waits for text input, returns it
(define (ask question)
  (begin (display question)
         (read)))
; greet : nothing -> prints output, waits for text input, prints output
(define (greet)
  (local [(define name (ask "What is your name?"))]
    (printf "Hello, ~a!" name)))
(check-expect
  (with-io-strings "Steve" greet)
  "What is your name?Hello, Steve!")}|
}

@; @include-section{worlds.scrbl}

@; @include-section{universes.scrbl}